Manuscript Number: HYDROL10488R2
Title: WEATHERING OF PLAGIOCLASE ACROSS VARIABLE FLOW AND SOLUTE TRANSPORT REGIMES
Article Type: Research Paper
Keywords: gravity flow fracture; capillary flow microfracture; advective transport; diffusive transport; open system; semi-open system; weathering; plagioclase; hydraulic diffusivity
Corresponding Author: Dr. Fernando A.L. Pacheco,
Corresponding Author's Institution: Trás-os-Montes e Alto Douro Univ. First Author: Fernando A.L. Pacheco
Fernando A.L. Pacheco
Department of Geology and Centre for Chemistry Trás-os-Montes and Alto Douro University
5000 Vila Real Portugal
Professor Laurent Charlet Editor
Journal of Hydrology November 7, 2011
HYDROL10488R1: Editor decision - revise (minor revision)
Dear Prof. Charlet,
We are very pleased with your decision regarding our manuscript. We addressed the very minor corrections pointed out by reviewer #1, related to incorrect citations to borehole numbers, as can be consulted in the DOC file Revision_Notes. The revised version of the manuscript is given in the DOC file Revision_Changes_Marked (an annotated manuscript showing the very minor changes relative to the original manuscript (HYDROL10488R1), affecting just three lines of that ms. The final version of the manuscript is given in the DOC file Manuscript_Final. Tables and corresponding legends are provided as a separate DOC file, the same happening with the figure captions. Figures as provided as TIFF format images, with resolution 600 or 1200 DPI.
We hope you are satisfied with the revised version, and that the paper can be published soon.
Kind regards,
Revision Notes 7 November 2011
Revision Notes
Reviewer #1
Correct the "Nr" in text.
The reviewer is right
Line178: Nr23 ? Nr22 Changed Nr 23 to Nr 22
Line183 and Line192: Nr4, 18, and 19 ? Nr4, 19 and 20 (?)
Changed Nr 4, 18 and 19 to Nr 4, 19 and 20
Line402: Nr10, 11, 16 and 25 ? Nr11, 12, 17 and 26
RESEARCH HIGHLIGHTS
Weathering studies based on the microsystem concept were initiated by Korzhinskii (1959), and recently developed by Meunier et al. (2007). Basically, these studies associate
precipitation of specific secondary products (e.g. smectite, halloysite) to specific locations in a fractured rock mass (e.g. capillary flow micro fractures, gravity-flow fractures) where specific flow and transport regimes prevail (e.g. chemical diffusion, advection). The scale of these studies was the thin section scale. The first research highlight of this paper is the extrapolation of the weathering analysis based on the microsystem concept to the macroscopic scale, applying it to fault zones underlying a soil plus saprolite cover.
White and Brantley (2003) showed that factor time explains the discrepancy reported in many studies between weathering rates estimated in the field and in the laboratory. The field weathering rates reported in that study were mostly solid-state weathering rates based on the differences between elemental, isotopic and mineral compositions measured in present-day regoliths and in the assumed protolith. In these cases, variable time (t) represents the entire time span of a weathering episode, commonly on the order of thousands to million years. These rates are not directly reconcilable with solute-flux rates, also estimated in field studies, that stand for contemporary weathering during groundwater percolation in the fractured rocks, in which case t (travel time) is just a portion of that time window, frequently ranging from years to decades. The second highlight of this paper is a demonstration that reconciliation between solid-state and solute-flux weathering rates can be accomplished if the change in aquifer properties (hydraulic diffusivity, which is a proportion of the ratio between hydraulic conductivity and effective porosity) over the geologic times is accounted for in the weathering models.
References
Korzhinskii, D.S. (1959). Physicochemical Basis of the Analysis of the Paragenesis of Minerals (translation). Consultant Bureau, New York, 143 pp.
Meunier, A., Sardini, P., Robinet, J.C., Prêt, D. (2007). The petrography of weathering processes: facts and outlooks. Clay Miner., 42: 415–435.
White, A.F., Brantley, S.L. (2003). The effect of time on the weathering of silicate minerals: why do weathering rates differ in the laboratory and field? Chem. Geol., 202: 479–506.
Text and References 7 November 2011
Text and References 1
WEATHERING OF PLAGIOCLASE ACROSS 1
VARIABLE FLOW AND SOLUTE TRANSPORT REGIMES 2
3 4
Fernando A. L. Pacheco,a 5
Department of Geology & Centre for Chemistry, 6
Trás-os-Montes and Alto Douro University, Ap 1013, 5000 Vila Real 7
Portugal 8
9
Cornelis H. Van der Weijden,b 10
Department of Earth Sciences—Geochemistry, Faculty of Geosciences, 11
Utrecht University, P.O. Box 80.021 3508 TA Utrecht, 12
The Netherlands 13
14
Key-words: gravity flow fracture, capillary flow microfracture, advective transport, diffusive 15
transport, open system, semi-open system, weathering, plagioclase, hydraulic diffusivity. 16
17
a Corresponding author, e-mail [email protected], Fax. +351 259 350480 b e-mail [email protected], Fax +31 30 2535302
Text and References 2 Abstract
18 19
The study area is situated in a fault zone with fractured granites and metassediments. In a 20
conceptual model, infiltrating water first passes the bedrock cover of soil and saprolite and 21
then partly enters the fractures. Weathering reactions of minerals occur in small pores and 22
fissures in the bedrock cover zone to continue in the larger fractures. Pumping tests were 23
carried out in a number of boreholes to measure the drawdown as a function of pumping 24
time. From the results, values of transmissivity (T) could be derived. In combination with the 25
storage coefficient (S) for similar fault zones, the hydraulic diffusivity (D = T/S) could be 26
computed. 27
Water samples, collected from the boreholes, represent fluid packets with a history of 28
weathering reactions in the bedrock cover and in the larger fractures. The major element 29
composition of these samples was used by means of the SiB mass balance algorithm 30
(Pacheco and Van der Weijden, 1996) to calculate the moles·L1 of dissolved plagioclase 31
(oligoclase with An ≈ 0.20) and the moles·L1 of secondary phases (gibbsite, halloysite, 32
smectite) precipitated along the flow paths of the samples. These results were then used to 33
calculate the net dissolved silica concentrations ([H4SiO04]) related to dissolution of
34
plagioclase followed by precipitation of each of the secondary phases. An interpretation of a 35
plot of each of these [H4SiO04]’s versus D is that at D < 0.7 m2·s–1, dissolution of plagioclase 36
is followed by precipitation of halloysite in the large fractures of the fault zone (open 37
system), whereas at D ≥ 0.7 m2·s–1 precipitation of both halloysite and smectite occurs in the 38
rock matrix with small fissures and pores (semi-open system). Before being pumped, the 39
percolating fluids travelled 0.01 to 13.7 years. During these periods, plagioclase weathered at 40
rates (WPl) of 10–(12.9±1.1) moles·m–2·s–1, which are approximately 2.2 orders of magnitude
41
higher than solid-state weathering rates reported in various field studies. In this study, it is 42
Text and References 3 suggested that part of the apparent discrepancy between the results is due to changes in 43
hydraulic diffusivity of the weathering environments occurring over the geologic times. 44
45 46
Text and References 4 1. Introduction
47 48
Weathering of rock forming minerals can be investigated in terms of microsystems, a concept 49
introduced by Korzhinskii (1959). These systems are located at each water–mineral contact 50
and are composed of the primary and secondary minerals and local solutions. They belong to 51
three categories: closed, semi-open and completely open (Meunier et al., 2007). In crystalline 52
rocks, closed microsystems are located at primary porosity and microfracture dead ends, 53
semi-open at microfractures and completely open at small fractures. 54
Groundwater flow and solute transport regimes within a microsystem depend on their 55
category: in a closed system, fluids are stagnant and therefore solute transport is limited; in a 56
semi-open system, water moves by capillarity and solute transport occurs by chemical 57
diffusion; and finally, in an open system, groundwater flow is controlled by gravity and 58
solute transport occurs by advection. The prevailing regimes determine the secondary 59
products derived from alteration of the parent minerals (Baynes and Dearman, 1978; Sausse 60
et al., 2001): in microfractures, where the leaching of solutes is less intense because transport 61
is associated to chemical gradients, the most common secondary products are smectites and 62
vermiculites; in fractures the leaching is intensified because solutes are carried away by the 63
flowing water and, as a consequence allophone and/or halloysite are formed; if velocity 64
and(or) the volume of flow increase, the leaching of (earth)alkaline elements and silica may 65
become extreme, resulting in formation of gibbsite. 66
Dissolution of parent minerals and precipitation of secondary products within microsystems 67
is accompanied by the alteration of rock porosity and permeability. In the process of chemical 68
weathering, water increasingly invades the microsystems developing secondary porosity that, 69
regardless of the quantities of the newly formed phases, generally increases because the 70
alteration products never completely replace the dissolved volume (Putnis, 2002; Meunier et 71
Text and References 5
al., 2007). The increase in porosity changes the geometrical parameters of the pathways. 72
Microfractures tend to widen and connect to other microsystems, increasing permeability 73
(Sardini et al., 2001; Sausse et al., 2001). In later stages of weathering, associated to the 74
development of saprolites, the microstructure of the rocks tends to disappear by collapsing or 75
sliding under gravity, causing a substantial decrease in permeability, but not in porosity 76
(Wright and Burgess, 1992). 77
Weathering studies based on the microsystem concept have been performed at the 78
microscopic scale (Meunier and Velde, 1979; Hochella and Banfield, 1995; Sausse et al, 79
2001; Meunier et al., 2007), using thin sections of rock samples as investigating materials. 80
The main goal of this paper is to extrapolate the analysis of microsystems to the macroscopic 81
scale, applying it to fault zones underlying a soil plus saprolite cover. The selection of fault 82
zones to study the macroscopic behaviour of microsystems is adequate, because fault zones 83
are complex fractured environments composed of large gravity-flow fractures in hydraulic 84
connection with a network of small gravity-flow fractures and capillary-flow microfractures 85
(rock matrix). At the scale of a fault zone, the evolution of pathway parameters can be 86
described by hydraulic diffusivity, which is the ratio of hydraulic transmissivity (a proportion 87
of permeability) to the storage capacity (a function of porosity). The aim is to verify if 88
changes in the hydraulic diffusivity have an effect on the type of secondary minerals formed 89
by weathering of plagioclase. 90
A complementary objective of this study is to calculate plagioclase weathering rates, taking 91
into account particularities of the fractured rocks relative to groundwater travel time and area 92
exposed to groundwater. The aim is to verify if the relation between weathering rates and 93
groundwater travel time can also be described in terms of progressive hydraulic diffusivity 94
changes. 95
Text and References 6 2. Study Area
96 97
The region of Vila Pouca de Aguiar is located in the North of Portugal and occupies an area 98
of approximately 437 km2 (Figure 1). It is characterized morphologically by large-scale 99
tectonic valleys, associated to the so-called Vila Real fault, surrounded by the Alvão (to the 100
West) and Padrela (to the East) mountains. Altitudes range from a minimum of 200 m in the 101
Northern valleys and a maximum of 1200 m up in the mountains. 102
Climate in the area is temperate with wet–cold and dry–warm alternating seasons. The annual 103
precipitations range from 900 mm·yr1 in the Northeast to 1900 mm·yr1 in the Southwest of 104
the region (Figure 2), being influenced by the topography. 105
The geology of the area is characterized by Hercynian (syn- to post-tectonic) granites that 106
intruded Palaeozoic (Cambrian to Devonian) metassediments and were covered by 107
Quaternary alluvial and terrace deposits along the Vila Real fault (Figure 2). In the Southeast 108
part of the area, occupied by an extensive outcrop of metassediments, a geological structure 109
was defined, in which tectonic laminae are folded and separated by major thrusts (Pacheco et 110
al., 1997). 111
The Vila Pouca de Aguiar granites are composed of (in weight %) quartz (31.5), plagioclase 112
(35.3), K-feldspar (26.7), biotite (4.4) and muscovite (1.4), with minor amounts of apatite and 113
ilmenite. The plagioclase composition ranges from albite-An8 to andesine-An37 (Pacheco et 114
al., 1999). The metassediments, which include greywackes, phyllites, quartzites and graphitic 115
slates, contain quartz, muscovite and smaller amounts of biotite, K-feldspar and albite-116 oligoclase (Pacheco, 1995). 117 118 119 120
Text and References 7 3. Hydrogeologic Setting
121 122
3.1. Conceptual Flow and Weathering Models 123
The fate of infiltrated rainwater in a fractured massif comprises lateral flow through the 124
bedrock cover (soil and saprolite) and recharge. Recharge includes refilling of the rock 125
matrix, composed of a network of small gravity flow fractures and capillary flow 126
microfractures, and of fault zones composed of large gravity-flow fractures and crushed rock. 127
Figure 3 illustrates the lateral flow of shallow groundwater through a saprolite horizon with a 128
thickness z0 (m), succeeded by its infiltration into a couple of sub-vertical fault zones. It also 129
shows paths of groundwater through the rock matrix until they reach the same fault zones. 130
Similar models of groundwater flow through fractured rocks towards fault zones have been 131
adopted by Herbert et al. (1992) and Verweij (1995). 132
The saprolite horizon is assumed an open system given the high effective porosity and 133
sometimes (granite saprolites) also high hydraulic conductivity of the disaggregated 134
materials. Depending on precipitation (P, mm), which determines the volume of flow, 135
secondary products expected to result from weathering of plagioclase are halloysite (lower 136
precipitations) or gibbsite (higher precipitations). In Portugal, gibbsite is the most common 137
secondary product for annual precipitations higher than about 1000 mm·y1, whereas for 138
precipitations lower than this limit the dominant product is halloysite (Martins et al., 1995). 139
The fault zones are also assumed open systems because large gravity-flow fractures dominate 140
in these environments. However, no extreme flushing conditions are expected to occur given 141
the compaction of these zones caused by the burden of overlying rocks. For that reason, 142
plagioclase weathering is expected to produce halloysite. Finally, the rock matrix is assumed 143
a semi-open system given the dominance of small gravity-flow fractures and capillary-flow 144
Text and References 8 microfractures in this environment. In this case, halloysite and smectite are both assumed 145
products of plagioclase weathering (Meunier et al., 2007). 146
147
3.2. Boreholes 148
In the past three decades, about 50 to 60 wells were drilled in the region of Vila Pouca de 149
Aguiar to be used as sources of drinking water for public supply. The location of 31 of the 150
most productive boreholes is plotted in Figure 1 (labelled circles) and the information on 151
name (Id), intersected rock type (granite or metassediment), depth (z, m), slope of the 152
topographic surface (d in %, sampled from the Digital Elevation Model of Figure 1) and local 153
precipitation (P in mm·y1, sampled from Figure 2), is depicted in Table 1. The associated 154
yields are illustrated in Figure 2 and can be justified by a combination of factors 155
encompassing the high annual precipitation, the abundance of quartzites within the 156
metassediments, and especially the proximity of the borehole to the Vila Real fault, to folds 157
and thrusts, or to the contact between granites and metamorphic rocks, places where the 158
fracture sets are usually more dense and interconnected. A number of studies reported the 159
proximity of boreholes to densely fractured regions as an important cause for observed high 160
yields (Siddiqui and Parizek, 1971; Mabee, 1999; Neves and Morales, 2007; among others). 161
During the drillings, several gravity-flow fractures were intersected by the boreholes 162
releasing variable amounts of water. The depth of intersection of fracture i (zi, m) and its 163
corresponding yield (Qi, m3·h–1) are listed in Table 1 for the above mentioned 31 boreholes. 164
Depths of intersection and yields were reported by the well drillers. For the first fracture, the 165
depth was set when water started flowing upward through the borehole emerging naturally at 166
the surface. For the other fractures of the same borehole, depths were set when an increase in 167
the natural outflow was noted. Yields were determined as the rate of water that can be 168
airlifted on a continuous short-term (generally tens of minutes) basis. These approaches for 169
Text and References 9 estimating fractured aquifer depths and yields are common and have been used in many other 170
studies (e.g., Moore et al., 2002). The average depth of the gravity flow fractures (zmed, m) 171
was estimated by weighting the depth of each fracture according to its yield: 172 j j j Q Q z z n 1 j n 1 j med (1) 173
where n is the number of fractures intersected by the borehole. 174
175
3.3. Pumping Tests and Hydraulic Parameters 176
Pumping tests were conducted on 14 of the 31 boreholes, lasting in all cases but one from 177
1140 to 2760 minutes. The exception was the test conducted on borehole Nr 22 (Cevivas) 178
which ran for just 330 minutes. The duration of the tests and their pumping rates and 179
hydrostatic levels are listed in Table 2. 180
In all cases, the scatter points drawdown (s, m) versus time (t, min) in a semilogarithmic plot 181
could be fitted with a straight line, after the first 100 minutes of pumping. In three cases (tests 182
to boreholes Nr 4, 19 and 20), the points could be linearly fitted with different two slopes, the 183
first after 10 minutes of pumping and the second after 100 minutes of pumping. Figure 4 184
illustrates the test to borehole Nr 4. The fitting of (s,t) points to one straight line in a 185
semilogarithmic plot is conform with type curves derived for porous media aquifers, and for 186
that reason aquifer transmissivities (T, m2·s–1) based on the straight lines fitted to the after-10 187
and/or after-100 minutes period(s) were calculated using a classical porous media model (the 188 Cooper-Jacob, 1946, formula): 189 T 2.3Q 4s (2) 190
where Q (m3·s–1) is the pumping rate and s (m) the change in s corresponding to a log cycle 191
of time. The two fittings reported for boreholes Nr 4, 19 and 20 can be attributed to 192
heterogeneity of the tested rocks: boreholes drain several gravity-flow fractures, with 193
Text and References 10 different hydraulic heads, that apparently reacted to pumping at two different moments 194
during the test. The influence of rock heterogeneity on pumping test results and interpretation 195
has been quoted by Dagan (1986), Newman (1990), Belcher et al. (2001), Kollet and Zlotnik 196
(2005), among other workers. 197
The hydraulic diffusivity (D, m2·s–1) is the quotient of transmissivity (T, m2·s–1) and storage 198
coefficient (S, dimensionless): 199
D = T/S (3)
200
Fractured rocks and regoliths derived therefrom span a wide range of D values, as depicted in 201
Table 3. The hydraulic conductivity (K, m·s–1) is the quotient of transmissivity and the 202
drainable thickness (b, m): 203
K = T/b (4)
204
The drained thicknesses of the tested boreholes are listed in Table 2 and were estimated by 205
multiplying the number of intersected fractures (Table 1) by the length of the screened tubes 206
casing the borehole around those fractures (in general 6 m, but 3 or 1 m for boreholes 207
Cevivas and Santa Marta, respectively). Effective porosity (ne, dimensionless) is the volume 208
of connected voids in a sample divided by the total volume of the sample, and can be derived 209
from the storage coefficient (Domenico and Schwartz, 1998): 210 w w e b S n (5) 211
where w is the specific weight of water (999.1 kg·m–3 at T = 15ºC) and w the 212
compressibility coefficient of water (4.7×10-9 m2·kg–1). 213
4. Sampling and AnalysisSamples of borehole water were collected within the limits of the 214
Vila Pouca de Aguiar municipality from January 2006 to April 2006. The sampling sites are 215
shown in Figure 1. At the sampling site, electrical conductivity (EC), temperature (T, ºC) and 216
pH were measured and a water sample was filtered through a 0.4 m membrane filter. The 217
Text and References 11 filtered water was split into 2 portions; one was stored for analysis in the home laboratory, 218
acidified to pH 2 using pure nitric acid, and the other for analysis of alkalinity in the field 219
laboratory, within 24 hours using the Gran plot method. In the home laboratory, ICP-OES 220
was used for analysis of Si. The analytical results are shown in Table 4. 221
222
5. Weathering Reactions and Rates 223
224
Not all minerals present in the granites or in the metassediments are important as weathering 225
reactants. In general, the abundance of plagioclase in granites is significant and therefore the 226
relative contribution of plagioclase weathering to groundwater composition is expected to be 227
expressive. However, because plagioclase has a low resistance to weathering, when compared 228
to other minerals composing these rocks (e.g. quartz, biotite, muscovite; Bland and Rolls, 229
1998; Appelo and Postma, 2005), this relative contribution may not be just expressive but 230
dominant. For example, in granite areas of central Portugal, Pacheco and Van der Weijden 231
(1996) and Van der Weijden and Pacheco (2006) reported contributions greater than 90%. In 232
the Vila Pouca de Aguiar granites, this dominance was also recognised by Pacheco et al. 233
(1999), which reported contributions between 95 and 98%, and of 99% when weathering was 234
associated to fault zones. In the metassediments, plagioclase may be present solely in small 235
amounts. However, weathering of plagioclase may also contribute dominantly to ground 236
water composition, because the surrounding minerals (e.g. quartz, chlorite) are usually much 237
less reactive. For example, in an area with metassediments where groundwater chemistry was 238
explained by weathering of plagioclase (albite) and chlorite (Pacheco and Alencoão, 2006 ), 239
the abundance of chlorite is about twice the abundance of albite, but on average the relative 240
contribution of albite weathering to groundwater composition is greater than 90%. For the 241
Text and References 12 reasons described above, it is assumed in this case that plagioclase weathering explains the 242
natural composition of ground water, namely the concentrations of silica and bicarbonate. 243
244
5.1. Mass Balances 245
Reactions of plagioclase producing gibbsite, halloysite, or smectite are listed in Table 5. The 246
contribution of each reaction to groundwater composition can be accomplished by solving a 247
set of mole balance equations based on the reaction stoichiometries, namely on the 248
bicarbonate to dissolved silica ratio (r) which is a key parameter in distinguishing between 249
different reactions (Garrels, 1967; Pacheco and Van der Weijden, 1996). For the reactions 250
listed in Table 5, the ratios are: 251 rGibb 1x 3x (6a) 252 x 1 x 1 2 1 Hal r (6b) 253 x 65 . 5 95 . 0 ) x 1 ( 6 . 0 Sm r (6c) 254
where x is the anorthite content of plagioclase (0 ≤ x ≤ 1), and the mole balance equations 255 are: 256
Sm 0 4 4 Sm Hal 0 4 4 Hal Gibb 0 4 4 Gibb3 [HSiO ] [HSiO ] [HSiO ]
HCO r r r (7a)
257
[H4SiO04][H4SiO04]Gibb[H4SiO04]Hal[H4SiO40]Sm (7b) 258
where square brackets represent dissolved concentrations. The unknowns of the system are x 259
and the partial concentrations of dissolved silica. An average value for x can be obtained 260
from chemical analysis of granite and metassediment plagioclases (in the Vila Pouca de 261
Aguiar region, plagioclase is faithfully represented by an oligoclase, meaning that x 0.2). 262
However, to become a determined system, set 7a,b must be completed with an additional 263
equation. 264
Text and References 13 According to the conceptual flow and weathering models illustrated in Figure 3, the
265
concentration derived from reaction RGibb (
[H4SiO04
]Gibb) is released when shallow 266
groundwater moves along the saprolite path, whereas the concentrations derived from RHal 267
(
[H4SiO04]Hal)and RSm (
[H4SiO04]Sm) are released or consumed along the fault zone plus rock 268
matrix paths. In the first case, the movement of water is sideways and water-mineral 269
interactions occur within a relatively narrow depth range (from ground surface to z0). Solute 270
concentrations will be higher or lower depending on whether the length of the saprolite flow 271
path is longer (lsp in Figure 3) or shorter (ssp), but they will be independent of circulation 272
depth (z). In the second case, water moves laterally as well as vertically, being sampled in the 273
boreholes at a depth z = zmed. The value of zmed varies from place to place (e.g., from 18 to 96 274
meters in the study area, Table 1). If dissolution of plagioclase and transport of silica along 275
the fault zone plus rock matrix paths are assumed continuous, then a range in circulation 276
depths are expected to follow a concomitant range in the dissolved silica concentrations, the 277
overall result being a regression between
[H4SiO4 0
]Hal[H4SiO4 0
]Smand zmed. This relationship 278
is not universal but can be defined for each particular case. With this relation added to 279
Equations 7a,b, the set becomes determined, and a unique solution can be found for its 280
unknowns. 281
282
5.2. Moles of Dissolved Plagioclase 283
The moles of plagioclase ([Pl]) dissolved along the saprolite, fault zone and rock matrix flow 284
paths are obtained by multiplying the concentrations of silica derived from reactions RGibb, 285
RHal and RSm by the concomitant plagioclase to silica ratios (), as deduced from Table 5: 286
[Pl]Gibb[H4SiO04]GibbHal[H4SiO40]HalSm[H4SiO04]Sm (8) 287 where 288 x Gibb β 31 (9a) 289
Text and References 14 ) 1 ( 2 1 x Hal β (9b) 290 x . . . Sm β 095165565 (9c) 291
The first term of the right-hand side of Equation 8 represents the number of moles of 292
plagioclase dissolved along the saprolite path ([Pl]0) whereas the sum of the second and third 293
terms represent the moles dissolved along the fault zone plus rock matrix paths. 294
295
5.3. Plagioclase Weathering Rates 296
Weathering rates of plagioclase can be estimated using: 297 WPl [Pl][Pl]0 t Vr APl (10) 298
where WPl (mol·m–2·s–1) is the rate and [Pl][Pl]0 (mol·m–3) its concomitant dissolved 299
concentration, t (s) is the average groundwater travel time of water packets flowing through 300
the fault zone and through the rock matrix, Vr (m3·s–1) is the volume of water entering the
301
fault zone in a unit time, and APl (m2·s–1) is the surface area of plagioclase in contact with that 302
volume of aquifer water. An equation for APl was developed by Pacheco and Alencoão 303
(2006), describing the area of fracture surfaces in contact with aquifer water in unit time: 304 APl2PlVr wgne 12wK (11) 305
where Pl is the proportion of plagioclase in the rock, w (1.14 103 kg·s
–1
·m–1 at T = 15ºC) 306
is the dynamic viscosity of water, and g is the acceleration of gravity (9.81 m·s–2). Replacing 307
this equation in Equation 10 and rearranging gives: 308 WPl[Pl][Pl]0 2tPl 12wK wgne (12a) 309
Further substitution of Equation 5 in Equation 12a gives 310 WPl[Pl][Pl]0 2tPl Cw D (12b) 311
Text and References 15 where 312 6 10 56 . 2 12 g C w w w s1/2 (at T = 15ºC) (12c) 313
Equation 12b describes the weathering rate as an explicit function of hydraulic diffusivity. 314
315
6. Groundwater Travel Times 316
317
The average groundwater travel time is the length of the flow path divided by the mean 318
velocity of water along that path. According to the conceptual flow model (Figure 3), water 319
will move towards the fault zones following two distinct paths: the fault zone path or the rock 320
matrix path. The length of the fault zone path can be equated to zmedz0, assuming that fault 321
zones are almost vertical, where zmed is the average depth to the conductive fractures 322
(Equation 1) and z0 is the thickness of the saprolite layer. The length of the rock matrix path 323
is difficult to assess, because it depends on the average distance between the recharge areas 324
and the fault zones as well as on the tortuosity of the path, which are unknowns. In this study, 325
the length of the flow path was approached to zmedz0.This provides minimum values for 326
flow path lengths and consequently to groundwater travel times, maximizing weathering rates 327
(Equation 12). Difficulties in determining the rock matrix flow path length were also 328
encountered and not resolved by Meunier et al. (2007). The mean velocity of water along the 329
flow path (v, m·s–1) is (Domenico and Schwartz, 1998): 330 e n Ki v (13) 331
where i is the hydraulic gradient (dimensionless). Although hydraulic gradients are 332
influenced by recharge and discharge rates as well as by groundwater withdrawal, they follow 333
primarily the local relief (Appelo and Postma, 2005), being determined by the slope of the 334
topographic surface (d, dimensionless). Hence, the travel time of groundwater moving from 335
Text and References 16 the bottom of the saprolite layer till the average depth of conductive fractures can be written 336 as: 337 t ( zmedz0)ne Kd (14) 338
It should be mentioned that K and ne were estimated from results of long-term pumping tests, 339
and therefore correspond to an average of the hydraulic conductivities and effective porosities 340
of the fault zone plus rock matrix. 341
342
7. Results and Discussion 343
344
7.1. Hydraulic Tests 345
The hydraulic test of borehole Bornes 1 (Figure 4) ran at a pumping rate Q = 13.5×10–3 m3·h– 346
1
(Table 2). For the log cycle 10 to 100 minutes, the scatter points fit to a straight line with 347
drawdowns s10 = 2.5 and s100 = 7.8 m, respectively. The change in the drawdown is s1 = 5.3 348
m which results in T1 = 1.31×10 –4
m2·s–1 (Equation 2). For the next log cycle of time (100 to 349
1000 minutes) the scatter points fit to a less inclined straight line resulting in a smaller s2 = 350
2.3 m and a larger transmissivity T2 = 3.05×10–4 m2·s–1. Even though the time spans linked to 351
T1 and T2 are usually different (Figure 4), the average fault zone transmissivities were 352
calculated using the arithmetic mean (T1+T2)/2. For the fault zone intersected by borehole 353
Bornes 1, Tmed = 2.18×10–4 m2·s–1. When considering the results of all pumping tests, Tmed = 354
(1.5±1.2)×10–4 m2·s–1. There were no observation wells around the tested boreholes so a 355
reasonable value had to be found for the storage coefficient. The choice was the average 356
storage coefficient of a 33.2 m thick fault zone estimated by Pacheco (2002): S = 357
(7.75±1.71)×10–5. Standing on this coefficient, the hydraulic diffusivities were calculated by 358
Equation 3 and compiled in Table 4. On average, D = 1.7±1.4 m2·s–1. These values lie in a 359
range of values observed for fractured weathered crystalline rocks (0.1–10 m2·s–1; Table 3) 360
Text and References 17 and are in agreement with results obtained in other studies (Talwani, 1981; Talwani and 361
Acree, 1984; Rastogi et al., 1986; Shapiro et al, 1997; Talwani and Cobb, 1999). 362
For borehole Bornes 1, b = 18 m and so Kmed = 12.1×10–6 m·s–1 (Equation 4). For the entire 363
set of tested wells Kmed = (8.8±5.7)×10 –6
m·s–1 (Table 2). Hydraulic conductivities decrease 364
steadily with zmed as shown in Figure 5. The scatter points in this figure can be fitted to an 365 exponential function: 366 med z e K 6 1 0.051 37 . 53 ) s m 10 ( , R2 = 0.7 (15) 367
This equation was used to estimate hydraulic conductivities for the non-tested boreholes 368
(Table 4). The effective porosity was based on the adopted storage coefficient (S = 7.75×10– 369
5
), and was estimated using Equation 5. The result was: ne = 0.50. 370
371
7.2. Weathering Reactions 372
The physicochemical parameters of the sampled waters are given in Table 4. During the 373
saprolite, fault zone and rock matrix paths, reaction of water with rock forming minerals 374
releases solutes and precipitates clay minerals, hydroxides, etc. One of the reaction-derived 375
solutes is silica (
H4SiO04) and for the studied borehole waters the concentrations of
H4SiO04 376
(mol·L–1) increase with depth (zmed, Figure 6). The sympathy between [
H4SiO40
] and zmed is 377
clear, in spite of the scatter bounded by two dashed lines. This scatter can be ascribed to the 378
contribution of saprolite weathering to [
H4SiO04], i.e. to weathering occurring before the 379
entrance of shallow ground water into the rock matrix and fault zones intersected by the 380
drilled wells, because this contribution is not supposed to depend on zmed . According to the 381
conceptual weathering model described before, the saprolite contribution to weathering is 382
referred to as
[H4SiO4 0
]Gibb. Since water moves laterally along the saprolite path (Figure 3), 383
interacting with minerals within a narrow range of circulation depths (0z0 meters), there is 384
Text and References 18 no reason why
[H4SiO04]Gibbshould depend on zmed. Instead,
[H4SiO04]Gibb should vary in 385
agreement with changes in the saprolite path length, as interpreted from a detailed analysis of 386
Figure 6. In this figure, we shall presume that
[H4SiO4 0
]Gibbis represented by a point in the 387
dotted line. This line marks the intersection between the lower dashed line and the X-axis (z0) 388
and stands for the average thickness of the saprolite layer (z0 = 16 m). If the path followed by 389
groundwater during the stage of saprolite weathering is the largest (lsp in Figure 3) then 390
[H4SiO04]Gibb is also the largest and represented by the intersection between the upper dashed 391
line and the dotted line. If this path is the shortest (ssp) then
[H4SiO04]Gibb will approach zero 392
and be represented by the intersection between the lower dashed line and the dotted line. In 393
general (values in Table 4), 394 [H4SiO04]Gibb = [H4SiO04]– 0.0043(zmed-z0) (16) 395
where 0.0043 is the slope of the dashed lines. 396
For the present case study, Equation 16 is required to convert Equations 7a,b into a set with a 397
unique solution for
[H4SiO04]Hal and
[H4SiO04]Sm. 398
The concentrations of silica derived from reactions RGibb, RHal and RSm (Table 5) are listed in 399
Table 4. In all cases
[H4SiO04]Gibband
[H4SiO04]Hal are positive, meaning that the corresponding 400
reactions invariably release silica to solution. With four exceptions (samples from drilled 401
wells Nr 11, 12, 17 and 26),
[H4SiO04]Smis negative, meaning that RSm generally consumes 402
silica from solution. These results are in keeping with the equations for the r ratios (Equations 403
6ac), because for any value of x rGibb and rHal are positive while rSm is positive only for x <
404
0.056 (albite). The exceptions represent cases where adoption of x = 0.2 is inadequate. For 405
the cases where RSm consumes silica, the reaction proceeds only if there is some silica 406
available in the solution, for example transported along the flow path from places where 407
Text and References 19 weathering of plagioclase has released this solute, i.e. from places where RGibb or RHal were 408
the prevailing reactions. 409
410
7.3. Weathering Reactions and Hydraulic Diffusivity 411
A plot of silica concentrations derived from reactions RHal and RSm (Table 5) versus hydraulic 412
diffusivity (D) is illustrated in Figure 7. For silica released by RHal, concentrations decrease as 413
D increases, according to a power function. When silica is consumed by RSm, this 414
consumption increases for increasing values of D, according to a linear function. 415
Fault zones are low diffusive environments because water tends to follow large gravity 416
fractures (the easiest routes). Contrarily, the rock matrix is potentially a high diffusive 417
environment because water has to move in a network of connected small gravity flow 418
fractures and capillary flow microfractures; the higher the connectivity the higher the 419
hydraulic diffusivity (Knudby and Carrera, 2006). According to this setting, fault zones must 420
be plotted to the left-hand side of Figure 7 and the rock matrix to the right-hand side of that 421
diagram. The boundary between the two environments must be drawn in keeping with the 422
conceptual weathering model described before. According to this model, reactions involving 423
plagioclase and occurring along fault zones (completely open systems) should be related to 424
precipitation of halloysite, and reactions occurring along the rock matrix (semi-open system) 425
to precipitation of smectite plus halloysite. In Figure 7, this scenario becomes more evident 426
for D 0.7 m–2·s–1: when D < 0.7 m–2·s–1, concentrations derived from RHal are high (average: 427
280 mmol·L–1) and concentrations related to RSm approach zero (average: 16 mmol·L –1
); for 428
D > 0.7 m–2·s–1, both reactions play a role in weathering of plagioclase (the average 429
concentrations are 158 mmol·L–1 for Rhal and 67 mmol·L–1 for Rsm), with a growing relative 430
importance of RSm. Although the limit of D = 0.7 m–2·s–1 may be considered arbitrary, it 431
Text and References 20 certainly can be used as provisional threshold for the separation between fault zone (open 432
system) and rock matrix (semi-open system) weathering. 433
An association can also be recognized between precipitation of halloysite and advective flow 434
along gravity fractures of the fault zones and rock matrix. This is illustrated in Figure 8. No 435
correlation exists between
[H4SiO4 0
]Gibb and zmed, but even was not expected as production of 436
gibbsite is ascribed to the saprolite path, not to the fault zone or rock matrix paths. A 437
correlation between the remainder of silica (
[H4SiO04]Hal +
[H4SiO04]Sm) and zmed was already 438
discussed (Figure 6) but Figure 8 shows that only
[H4SiO04
]Hal is correlated with depth. This is 439
a remarkable result because confirms that production of halloysite is indeed associated with 440
plagioclase weathering along the mean direction of flow, and that the same does not hold for 441
production of smectite. In keeping with these results, it may be concluded that production of 442
halloysite does occur along the walls of gravity-flow fractures, in fault zones or the rock 443
matrix. In these two media, solute transport is advective and therefore concentrations must 444
increase with increasing flow path lengths (zmed) justifying the observed correlation. 445
Conversely, production of smectite occurs at micro fractures adjacent to the gravity flow 446
fractures, where solute transport is diffusive, mediated by local gradients of chemical 447
potential and by the length of diffusion trajectories between the capillary and gravity-flow 448
fractures, and hence not necessarily dependent on the flow path lengths. 449
450
7.4. Groundwater Travel Times, Plagioclase Weathering Rates and Hydraulic Diffusivity 451
Groundwater travel times (t) calculated by Equation 14 are given in Table 4 and range from 452
0.01 to 13.7 years, being on average 2.2±3.5 yr. Plagioclase weathering rates (WPl) calculated 453
by Equation 12b are compiled in Table 4, with an average logarithmic value of 12.9±1.1. 454
These rates are plotted in Figure 9 (filled circles) along with field weathering rates reported in 455
Text and References 21 other studies (White and Brantley, 2003, and references therein; open circles), with an
456
average logarithmic value of 15.1±1.0. 457
Literature rates are on average 2.2 orders of magnitude lower than the rates calculated in this 458
study. In a first glance at Figure 9, it could be suggested that factor time is the sole 459
explanation for this discrepancy, but a thorough analysis of the various results reveals that 460
other factors are also important. The open circles are mostly solid-state weathering rates 461
based on the differences between elemental, isotopic and mineral compositions measured in 462
present-day regoliths and in the assumed protolith. In these cases, variable t represents the 463
entire time span of a weathering episode, commonly on the order of thousands to million 464
years. The filled circles are solute-flux rates that stand for contemporary weathering during 465
groundwater percolation in the fractured rocks, in which case t (travel time) is just a portion 466
of that time window, frequently ranging from years to decades. For the open circles, it is 467
inherently assumed that the weathering environment is initially a fresh rock (the protolith), 468
i.e. a medium with a very high hydraulic diffusivity (Table 3). For the filled circles, it is 469
recognised that many groundwater parcels have circulated through the fractured rocks over 470
the full time span of weathering, which had as consequence a progressive decrease in the 471
hydraulic diffusivity, till present day values that do not exceed 5 m2·s1 (Table 4). 472
The consequences for weathering rates of changing the hydraulic properties of the percolated 473
rocks are predicted by Equation 12b: the impact in the WPl values is proportional to the square 474
root of the change in the D values. To elucidate the full relationship between weathering rate, 475
hydraulic diffusivity and time, an equation describing the drop in D as a function of t has to 476
be defined. If, as working hypothesis, it is assumed that D decreases according to a power 477
function of t, then Equation 12b can be rewritten as: 478
/ 0 Pl 0 2 ] Pl [ ] Pl [ f w Pl C D t t W (17) 479Text and References 22 with t given in years and where f is the decreasing rate, converts years into seconds ( 480
=3.1536×107 s·yr1), and D0 is a proxy for the initial hydraulic diffusivity. The plot of 481
Equation 17 in Figure 9 is represented by the dotted lines, each one representing a different 482
D0. In this plot, it is assumed that [Pl] – [Pl]0 = 0.686 mol·m3 (the average value; Table 4), 483
that Pl = 0.353 (the proportion of plagioclase in the granites) and that f = 1. According to the 484
distribution of the D0 lines, most solid-state rates are indeed related to the weathering of a 485
protolith with 103 < D0 < 106 m2/s whereas solute-flux rates are related to a weathered 486
fractured rock with 10–2 < D0 < 102 m2/s. It is also clear that, in case the aquifers investigated 487
in this study were composed of average diffusivity protoliths (D0 = 105 m2·s1; Table 3), 488
instead of being composed of low diffusivity weathered fractured rocks (D0 100 m2·s1; 489
Table 4),the calculated rates would be 2.5 orders of magnitude higher (A-double arrowed 490
line in Figure 9). This difference may be defined as the impact on weathering rates resulting 491
from continuous recirculation of groundwater through the studied fault zones, over the 492
geologic times. Factor time adds up to the previous impact, being minimal for waters that 493
circulated through the system only a few days or months and maximum for waters with 494
residence times of several years or decades. In the studied region, the difference between 495
minimum and maximum rates spans 4 orders of magnitude (B-double arrowed line). In 496
summary, the past changes in hydraulic diffusivity till present day values account for 497
approximately 40% reduction in weathering rates (2.5 orders of magnitude), relatively to their 498
presumed original values, whereas travel time accounts for the remaining 60% (4 orders). 499
500
8. Conclusions 501
502
The drawdown (s) in boreholes (observed during pumping tests), plotted against log cycles of 503
pumping time (t) could be fitted by two straight lines, one for a short (10 to 100 minutes), the 504
Text and References 23 other for a longer (100 to 1000 minutes) period of pumping time. Using the latter results, the 505
average transmissivity (T) was estimated and, in combination with an average storage 506
coefficient (S) typical for regional fault zones, the hydraulic diffusivity (D = T/S) could be 507
calculated: on average D = 1.7±1.4 m2·s–1. Lower values of D are typical for fault zones with 508
large fractures, while higher values are indicative of connected networks of small to very 509
small fractures typical for the rock matrix. Applying the so-coined SiB mass balance 510
equations (Pacheco and Van der Weijden, 1996) to the water composition, the contribution of 511
dissolved plagioclase to the water chemistry as well as the amounts of secondary precipitates 512
(gibbsite, halloysite, smectite) withdrawn from these waters can be calculated. Next, the net 513
release of dissolved silica from plagioclase dissolution in combination with precipitation of 514
gibbsite and halloysite and its consumption by smectite formation can be computed. The net 515
release of dissolved silica derived from plagioclase dissolution in combination with halloysite 516
precipitation correlates with the average depth (zmed) of the productive sectors of the drilled
517
wells, strongly suggesting halloysite precipitation along fractures in the bedrock. 518
Combinations of dissolved silica concentrations derived from plagioclase dissolution with 519
precipitation of gibbsite or smectite lack a correlation with zmed. The relation between the net
520
dissolved silica concentrations and D suggests that for D < 0.7 m2·s–1 the weathering of 521
plagioclase is largely followed by precipitation of halloysite and takes place in the completely 522
open system of the fault zone, whereas at higher values reactions occur also in the rock 523
matrix with precipitation of both halloysite and smectite. Groundwater travel times are on 524
average 2.2±3.5 year. Weathering rates (WPl’s) are calculated from the moles of plagioclase
525
dissolved along the flow path per unit of area of plagioclase exposed to the fluid and travel 526
time t. The average (WPl) = 10–(12.9±1.1) mol·m2·s–1 is 2.2 orders of magnitude higher than
527
solid-state weathering rates reported in various studies. The explanations for the apparent 528
discrepancy are partly the past changes in hydraulic diffusivity of the weathering 529
Text and References 24 environments, from approximately 105 m2/s in the protoliths to 0.11 m2/s in present-day 530
weathered fractured rocks, and partly the duration of weathering. 531
Text and References 25
References
533 534
Abrisqueta, J. M., Plana, V., Ruiz-Canales, A., Ruiz-Sánchez, M. C, 2006. Unsaturated hydraulic 535
conductivity of disturbed and undisturbed loam soil. Span. J. Agric. Res. 4(1), 91–96. 536
537
Appelo, C.A.J., Postma, D., 2005. Geochemistry, Groundwater and Pollution. A.A. Balkema, 538
Rotterdam, 649 pp. 539
540
Baynes J., Dearman, W.R., 1978. The microfabric of a chemically weathered granite. Bull. Eng. Geol. 541
Environ.18, 91100. doi: 10.1007/BF02635354 542
543
Belcher, W.R., Elliott, P.E., Geldon, A.L., 2001. Hydraulic-property estimates for use with a transient 544
ground-water flow model for the Death Valley regional ground-water flow system, Nevada and 545
California: U.S. Geological Survey Water-Resources Investigations Report 014210. url: 546
http://water.usgs.gov/pubs/wri/wri014210/. 547
548
Bland, W., Rolls, D., 1998. Weathering: and Introduction to the Scientific Principles. Arnold, London, 549
271p. 550
551
Cooper, H.H., Jacob, C.E., 1946. A generalized graphical method for evaluating the formation 552
constants and summarizing well field history. Transactions American Geophysical Union 27, 553
526534. 554
555
Dagan, G., 1986. Statistical theory of groundwater flow and transport: Pore to laboratory, laboratory 556
to formation, and formation to regional scale. Water Resour. Res. 22(9), 120S135S. 557
doi:10.1029/WR022i09Sp0120S 558
Text and References 26 Domenico, P.A., Schwartz, F.W., 1998. Physical and Chemical Hydrogeology (2nd ed.). John Wiley & 560
Sons Inc., New York, 506pp. doi: wiley.com/ 10.1002/ 1099-1034(200004/ 06)35 561
562
Garrels, R.M., 1967. Genesis of some ground waters from igneous rocks. In: Abelson, P.H. (Ed.), 563
Researches in Geochemistry, Wiley, New York, v. 2, pp. 405420. 564
565
Herbert, R., Barker, J.A., Kitching, R., 1992. New approaches to pumping test interpretation for dug 566
wells constructed on hard rocks. In: Burgess, E.P. (Ed.), Hydrogeology of crystalline basement 567
aquifers in Africa. Geological Society Special Publication, no. 66, pp. 221–242. doi: 10.1144/ 568
GSL.SP.1992.066.01.11. 569
570
Hochella, M.F., Banfield, J.F., 1995. Chemical weathering of silicates in nature: a microscopic 571
perspective with theoretical considerations. In: White, A.F. & Brantley, S.L. (Eds.), Chemical 572
Weathering Rates of Silicate Minerals. Reviews in Mineralogy 31, Mineralogical Society of America, 573
pp. 353–406. 574
575
Knudby, C., Carrera, J., 2006. On the use of apparent hydraulic diffusivity as an indicator of 576
connectivity. J. Hydrol. 329, 377–389. doi:10.1016/j.jhydrol.2006.02.026. 577
578
Kollet, S.J., Zlotnik, V.A., 2005. Influence of aquifer heterogeneity and return flow on pumping test 579
data interpretation. J. Hydrol. 300(14), 267–285. doi:10.1016/j.jhydrol.2004.06.011. 580
581
Korzhinskii, D.S., 1959. Physicochemical Basis of the Analysis of the Paragenesis of Minerals 582
(translation). Consultant Bureau, New York, 143 p. 583
584
Mabee, S.B., 1999. Factors influencing well productivity in glaciated metamorphic rocks. 585
Groundwater 37(1), 88–97. doi: 10.1111/j.1745-6584.1999.tb00961.x. 586
Text and References 27 Martins, A.A.A., Madeira, M.V., Refega, A.A.G., 1995. Influence of rainfall on properties of soils 587
developed on granite in Portugal. Arid Soil Res. Rehabil. 9, 353–366. doi: 588
10.1080/15324989509385904. 589
590
Meunier, A., Velde, B., 1979. Weathering mineral facies in altered granites: the importance of local 591
small-scale equilibria. Mineral. Mag, 43, 261–268. 592
593
Meunier, A., Sardini, P., Robinet, J.C., Prêt, D., 2007. The petrography of weathering processes: facts 594
and outlooks. Clay Miner. 42, 415–435. doi: 10.1180/claymin.2007.042.4.01 595
596
Moore, R.B., Schwarz, G.E., Clark Jr., S. F., Walsh, G.J., Degnan, J. R., 2002. Factors Related to 597
Well Yield in the Fractured-Bedrock Aquifer of New Hampshire. U.S. Geol. Survey Prof. Paper 1660, 598
22p. 599
600
Neuman, S.P., 1975. Analysis of pumping test data from anisotropic unconfined aquifers considering 601
delayed gravity response. Water Resour. Res. 11(2), 329342. 602
603
Neves, M.A., Morales, N., 2007. Structural control over well productivity in the Jundiaí River 604
Catchment, Southeastern Brazil. Annals of the Brazilian Academy of Sciences 79(2), 307–320. 605
606
Pacheco, F.A.L., 1995. Interacção Água-Rocha em Unidades do Grupo Peritransmontano (Serra da 607
Padrela-Vila Pouca de Aguiar). MSc Thesis, Coimbra University, Coimbra, 123p. 608
609
Pacheco, F.A.L., 2002. Response to pumping of wells in slopping fault-zone aquifers. J. Hydrol. 610
259(1–4), 116–135. doi: 10.1016/S0022-1694(01)00584-4 611
Text and References 28 Pacheco, F.A.L., Alencoão, A.M.P., 2006. Role of fractures in weathering of solid rocks: narrowing 613
the gap between experimental and natural weathering rates. J. Hydrol. 316, 248–265. 614
doi:10.1016/j.jhydrol.2005.05.003. 615
616
Pacheco, F.A.L., and Van der Weijden, C.H., 1996. Contributions of water-rock interactions to the 617
composition of groundwater in areas with a sizable anthropogenic input. A case study of the waters of 618
the Fundão area, central Portugal. Water Resour. Res. 32(12), 3553–3570. doi:10.1029/96WR01683. 619
620
Pacheco, F.A.L., Sousa, L.M.O., Sousa Oliveira, A., 1997. The Supply of a Small Town with 621
Groundwater from Thrusted Quartzites. In: Marinos, P.G. et al. (eds.), Engineering Geology and the 622
Environment, Balkema, Rotterdam, v. 2, pP. 14071412. 623
624
Pacheco, F.A.L., Sousa Oliveira, A., Van der Weijden, A.J., Van der Weijden, C.H., 1999. 625
Weathering, biomass production and groundwater chemistry in an area of dominant anthropogenic 626
influence, the Chaves-Vila Pouca de Aguiar region, north of Portugal. Water, Air Soil Poll. 115(1/4), 627
481512. doi: 10.1023/A:1005119121666. 628
629
Putnis A., 2002. Mineral replacement reactions: from macroscopic observations to microscopic 630
mechanisms. Mineral. Mag. 66, 689708. doi: 10.1180/0026461026650056 631
632
Rastogi, B. K., Rao, B. R., Rao, C. V. R. K., 1986. Microearthquake investigations near Sriramsagar 633
Reservoir, Andhra Pradesh State, India. Phys. Earth Planet. Inter. 44, 149–159. doi:10.1016/0031-634
9201(86)90041-5. 635
636
Roeloffs, E. A., 1997. Poroelastic techniques in the study of earthquake related hydrologic 637
phenomena. Adv. Geophys., 37, 135–195. doi:10.1016/S0065-2687(08)60270-8. 638
Text and References 29 Sardini, P., Sammartino, S. Tévissen, E., 2001. An image analysis contribution to the study of
640
transport properties of low-permeability crystalline rocks. Comput. Geosci. 27, 10511059. 641
doi:10.1016/S0098-3004(00)00157-6 . 642
643
Sausse, J., Jacquot, E., Leroy, J., Lespinasse M., 2001. Evolu tion of crack permeability during fluid– 644
rock interaction. Example of the Brézouard granite (Vosges, France). Tectonophysics 336, 199214. 645
doi:10.1016/S0040-1951(01)00102-0. 646
647
Shapiro, S. A., Huenges, E., Borm, G., 1997. Estimating the crust permeability 648
from fluid-injection-induced seismicity emission at the KTB site. Geophys. J. Int. 131, F15–F18. doi: 649
10.1111/j.1365-246X.1997.tb01215.x 650
651
Siddiqui, S.H., Parizek, R.R., 1971. Hydrogeologic factors influencing well yields in folded and 652
faulted carbonate rocks in central Pennsylvania. Water Resour. Res. 7(5), 2951312. 653
654
Talwani, P., 1981. Hydraulic diffusivity and reservoir induced seismicity. Final Tech. Rep., U.S. 655
Geol. Surv., Reston, Va, 48 p. 656
657
Talwani, P., Acree, S., 1984. Pore pressure diffusion and the mechanism of reservoir induced 658
seismicity. Pure Appl. Geophys. 122, 947–965. doi: 10.1007/BF00876395. 659
660
Talwani, P., Cobb, J., 1997. Induced seismicity studies at Bad Creek Reservoir, South Carolina. Final 661
Tech. Rep., U.S. Geol. Surv., Reston, Va, 89 p. 662
663
Van der Weijden, C.H., Pacheco, F.A.L., 2006. Hydrogeochemistry in the Vouga River basin (central 664
Portugal): pollution and chemical weathering. Appl. Geochem. 21, 580–613. 665
doi:10.1016/j.apgeochem.2005.12.006. 666
Text and References 30 Verweij, J.M., 1995. Analysis of Pumping Test Data from Hard Rock Aquifers. TNO Report, Institute 667
of Applied Geoscience, England, v. GG R-95-39(A), 65 p. 668
669
White, A.F., Brantley, S.L., 2003. The effect of time on the weathering of silicate minerals: why do 670
weathering rates differ in the laboratory and field? Chem. Geol. 202, 479–506. 671
doi:10.1016/j.chemgeo.2003.03.001. 672
673
Wright, E.P., Burgess, W.D. (1992). The Hydrology of Crystalline Basement Aquifers in Africa. 674
Special Publication 66, Geological Society, London. 264 p. doi: 10.1144/GSL.SP.1992.066.01.01. 675
Text and References 7 November 2011
Text and References 1
WEATHERING OF PLAGIOCLASE ACROSS 1
VARIABLE FLOW AND SOLUTE TRANSPORT REGIMES 2
3 4
Fernando A. L. Pacheco,a 5
Department of Geology & Centre for Chemistry, 6
Trás-os-Montes and Alto Douro University, Ap 1013, 5000 Vila Real 7
Portugal 8
9
Cornelis H. Van der Weijden,b 10
Department of Earth Sciences—Geochemistry, Faculty of Geosciences, 11
Utrecht University, P.O. Box 80.021 3508 TA Utrecht, 12
The Netherlands 13
14
Key-words: gravity flow fracture, capillary flow microfracture, advective transport, diffusive 15
transport, open system, semi-open system, weathering, plagioclase, hydraulic diffusivity. 16
17
a Corresponding author, e-mail [email protected], Fax. +351 259 350480 b e-mail [email protected], Fax +31 30 2535302
Text and References 2 Abstract
18 19
The study area is situated in a fault zone with fractured granites and metassediments. In a 20
conceptual model, infiltrating water first passes the bedrock cover of soil and saprolite and 21
then partly enters the fractures. Weathering reactions of minerals occur in small pores and 22
fissures in the bedrock cover zone to continue in the larger fractures. Pumping tests were 23
carried out in a number of boreholes to measure the drawdown as a function of pumping 24
time. From the results, values of transmissivity (T) could be derived. In combination with the 25
storage coefficient (S) for similar fault zones, the hydraulic diffusivity (D = T/S) could be 26
computed. 27
Water samples, collected from the boreholes, represent fluid packets with a history of 28
weathering reactions in the bedrock cover and in the larger fractures. The major element 29
composition of these samples was used by means of the SiB mass balance algorithm 30
(Pacheco and Van der Weijden, 1996) to calculate the moles·L1 of dissolved plagioclase 31
(oligoclase with An ≈ 0.20) and the moles·L1 of secondary phases (gibbsite, halloysite, 32
smectite) precipitated along the flow paths of the samples. These results were then used to 33
calculate the net dissolved silica concentrations ([H4SiO04]) related to dissolution of
34
plagioclase followed by precipitation of each of the secondary phases. An interpretation of a 35
plot of each of these [H4SiO04]’s versus D is that at D < 0.7 m2·s–1, dissolution of plagioclase 36
is followed by precipitation of halloysite in the large fractures of the fault zone (open 37
system), whereas at D ≥ 0.7 m2·s–1 precipitation of both halloysite and smectite occurs in the 38
rock matrix with small fissures and pores (semi-open system). Before being pumped, the 39
percolating fluids travelled 0.01 to 13.7 years. During these periods, plagioclase weathered at 40
rates (WPl) of 10–(12.9±1.1) moles·m–2·s–1, which are approximately 2.2 orders of magnitude
41
higher than solid-state weathering rates reported in various field studies. In this study, it is 42
Text and References 3 suggested that part of the apparent discrepancy between the results is due to changes in 43
hydraulic diffusivity of the weathering environments occurring over the geologic times. 44
45 46
Text and References 4 1. Introduction
47 48
Weathering of rock forming minerals can be investigated in terms of microsystems, a concept 49
introduced by Korzhinskii (1959). These systems are located at each water–mineral contact 50
and are composed of the primary and secondary minerals and local solutions. They belong to 51
three categories: closed, semi-open and completely open (Meunier et al., 2007). In crystalline 52
rocks, closed microsystems are located at primary porosity and microfracture dead ends, 53
semi-open at microfractures and completely open at small fractures. 54
Groundwater flow and solute transport regimes within a microsystem depend on their 55
category: in a closed system, fluids are stagnant and therefore solute transport is limited; in a 56
semi-open system, water moves by capillarity and solute transport occurs by chemical 57
diffusion; and finally, in an open system, groundwater flow is controlled by gravity and 58
solute transport occurs by advection. The prevailing regimes determine the secondary 59
products derived from alteration of the parent minerals (Baynes and Dearman, 1978; Sausse 60
et al., 2001): in microfractures, where the leaching of solutes is less intense because transport 61
is associated to chemical gradients, the most common secondary products are smectites and 62
vermiculites; in fractures the leaching is intensified because solutes are carried away by the 63
flowing water and, as a consequence allophone and/or halloysite are formed; if velocity 64
and(or) the volume of flow increase, the leaching of (earth)alkaline elements and silica may 65
become extreme, resulting in formation of gibbsite. 66
Dissolution of parent minerals and precipitation of secondary products within microsystems 67
is accompanied by the alteration of rock porosity and permeability. In the process of chemical 68
weathering, water increasingly invades the microsystems developing secondary porosity that, 69
regardless of the quantities of the newly formed phases, generally increases because the 70
alteration products never completely replace the dissolved volume (Putnis, 2002; Meunier et 71
